# Thread: Nested/Repeated Comparisons-What test should I use?

1. ## Nested/Repeated Comparisons-What test should I use?

Hi!

I am having trouble to decide what statistical test to use on my data because I have one factor grouped within another. Here is my experimental design: I have two cell lines, one infected with a bacteria (I) and one uninfected (U). I have split each of theses cell lines into two groups and treated one group with a chemical treatment (T) and one with a control treatment (C). So I have four groups of cells:

Infected-Treated
Infected-Control
Uninfected-Treated
Uninfected-Control

The uninfected cell line is like a control for the infected cell line.

I then ran a gene expression microarray with technical triplicates, so my dependent variable is continuous and normally distributed. First I need to compare T vs. C within each cell type (U and I)to determine the affect of the treatment on that cell type (can i use a unpaired T.Test for this?). Then ultimately I would like to compare the affect of the treatment between infected and uninfected cell types to determine how the bacterial infection alters genes response to the chemical treatment. Which statistical tests would be appropriate for this experimental design?

Any help for this thread would be greatly appreaciated!

Many Thanks

2. ## Re: Nested/Repeated Comparisons-What test should I use?

A repeated measures ANOVA sounds appropriate here.

3. ## Re: Nested/Repeated Comparisons-What test should I use?

Originally Posted by trinker
A repeated measures ANOVA sounds appropriate here.
Can't see where repetition comes from!
To me , this is a 2 by 2 factorial design and should be analyzed as a 2-way ANOVA with interaction. If the interaction term turns out to be insignificant, repeat the test without interaction, i.e. main effects model only.

4. ## Re: Nested/Repeated Comparisons-What test should I use?

Originally Posted by ASSAN
Can't see where repetition comes from!
To me , this is a 2 by 2 factorial design and should be analyzed as a 2-way ANOVA with interaction. If the interaction term turns out to be insignificant, repeat the test without interaction, i.e. main effects model only.
Technical replicates are measurements taken on the same thing essentially so you're really just doing a repeated measure of some cell.

@WhackyWolbachia - What software are you using for the analysis?

#### Posting Permissions

• You may not post new threads
• You may not post replies
• You may not post attachments
• You may not edit your posts